8374308: ImageBufferCache has no effect and can be removed by david-beaumont · Pull Request #29043 · openjdk/jdk

david-beaumont · 2026-01-05T17:19:50Z

Remove ineffective/unused ImageBufferCache class, and simplify callers / remove dead code.

I removed the release methods in the internal classes, but the public ModuleReader API method is still there (the override can go away though since the default implementation also tests for non-null, so removing the override has no risk).

I suspect there are no implementations of ModuleReader that implement release semantics after this change, so perhaps we could relax the documentation around it? Thoughts welcome.

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8374308: ImageBufferCache has no effect and can be removed (Bug - P4)

Reviewers

Alan Bateman (@AlanBateman - Reviewer)
Roger Riggs (@RogerRiggs - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/29043/head:pull/29043
$ git checkout pull/29043

Update a local copy of the PR:
$ git checkout pull/29043
$ git pull https://git.openjdk.org/jdk.git pull/29043/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 29043

View PR using the GUI difftool:
$ git pr show -t 29043

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/29043.diff

Using Webrev

Link to Webrev Comment

bridgekeeper · 2026-01-05T17:21:09Z

👋 Welcome back david-beaumont! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2026-01-05T17:21:40Z

@david-beaumont This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8374308: ImageBufferCache has no effect and can be removed

Reviewed-by: alanb, rriggs

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 22 new commits pushed to the master branch:

f1e0e0c: 8374544: Add SleepyCat diagnostics for all platforms
fbc59ac: 8374555: No need for visible input warning in s.s.u.Password when not reading from System.in
cdbc493: 8374190: Convert ConcurrentHashTable atomic lists to use Atomic
... and 19 more: https://git.openjdk.org/jdk/compare/6ae3e064352a56c5be140fba1ad6d040219432b0...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

As you do not have Committer status in this project an existing Committer must agree to sponsor your change. Possible candidates are the reviewers of this PR (@AlanBateman, @RogerRiggs) but any other Committer may sponsor as well.

➡️ To flag this PR as ready for integration with the above commit message, type /integrate in a new comment. (Afterwards, your sponsor types /sponsor in a new comment to perform the integration).

openjdk · 2026-01-05T17:22:31Z

@david-beaumont The following label will be automatically applied to this pull request:

core-libs

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

mlbridge · 2026-01-05T17:27:09Z

Webrevs

david-beaumont · 2026-01-05T17:23:54Z

src/java.base/share/classes/jdk/internal/jimage/BasicImageReader.java

        }
    }

+    private static ByteBuffer allocateBuffer(long size) {


The only bit of code from ImageBufferCache that was being used in existing code.

david-beaumont · 2026-01-05T17:24:54Z

src/java.base/share/classes/jdk/internal/jimage/BasicImageReader.java

+        return buffer != null ? getBufferBytes(buffer) : null;
    }

+    /**


Seemed worth stressing that this is a new buffer allocation (since there's no "release" option to callers anymore, it's a usage/lifetime change).

david-beaumont · 2026-01-05T17:28:40Z

src/java.base/share/classes/jdk/internal/jimage/ImageReader.java

-     * Returns the content of a resource node in a possibly cached byte buffer.
-     * Callers of this method must call {@link #releaseByteBuffer(ByteBuffer)}
-     * when they are finished with it.
+     * Returns the content of a resource node in a newly allocated byte buffer.


While the "ByteBuffer" instance is newly allocated, it's almost always a cheap slice of the memory mapped file. I'm not sure what we want as the best wording here, since:

We want people to know they don't need to worry about "releasing" it, but also

we don't want people to think it's using a newly allocated/copied buffer.

Happy to take suggestion on better wording here and for "getResourceBuffer" above.

david-beaumont · 2026-01-05T17:29:26Z

src/java.base/share/classes/jdk/internal/module/SystemModuleFinders.java

                    .map(reader::getResourceBuffer);
        }

-        @Override


Can be removed since ModuleReader has a default method with the null check in it.

AlanBateman · 2026-01-05T17:51:56Z

src/java.base/share/classes/jdk/internal/jimage/BasicImageReader.java

+        if (size < 0 || Integer.MAX_VALUE < size) {
+            throw new IndexOutOfBoundsException("size");
+        }
+        ByteBuffer result = ByteBuffer.allocateDirect((int) ((size + 0xFFF) & ~0xFFF));


Can you remind me why it allocates multiples of 4k?

Also, would be useful to know if using ByteByte.allocate (to allocate a heap buffer) would be okay here.

I can't, I've never seen this code before. No comments, no tests for why 4k is a good size. Probably just a "page size" heuristic. I'm just removing dead code and moving what's not dead out of the class to be deleted.

Changing details beyond that should probably be a different PR.

Actually, thinking about this more, it would have the benefit (in the situation where caching happens) of making buckets more likely to be reususeable, so if we're willing to assume that, and since the cache is going away, we could justify dropping the rounding and just allocating the exact size.

I can't see any reason to round up now so I think we should remove it. I don't mind if its this PR or a separate PR.

I'm interested to know if this can be changed to use allocate as it doesn't need a direct buffer here. FileChannel.read uses the thread-local buffer cache when called with a heap-buffer.

AlanBateman · 2026-01-05T17:52:39Z

src/java.base/share/classes/jdk/internal/jimage/BasicImageReader.java


+    private static ByteBuffer allocateBuffer(long size) {
+        if (size < 0 || Integer.MAX_VALUE < size) {
+            throw new IndexOutOfBoundsException("size");


I don't think it can happen but I assume IAE would be more appropriate here as size is not an index.

Only needs to check for the Integer.MAX_VALUE in any case, as ByteBuffer::allocateDirect throws on size < 0. It also uses IAE

Fair points. This is just moved code, but IOOBE is clearly not the right choice here.

In fact, taking a more holistic view, the size check is already done by the caller, so there's some nice simplification possible here.

You do need a negative check when using long because of underflow.
If longValue is 0xFFFFFFFF00000001 (negative, invalid) then '(int) longValue' is 1 (positive, valid).

https://docs.oracle.com/javase/specs/jls/se7/html/jls-5.html#jls-5.1.3

Or you could use a different style of check:

int intValue = (int) longValue;
if (intValue != longValue) { throw ... }

RogerRiggs · 2026-01-05T19:31:57Z

Please update the issue with the change in scope of the change.

AlanBateman · 2026-01-05T19:37:27Z

I suspect there are no implementations of ModuleReader that implement release semantics after this change, so perhaps we could relax the documentation around it? Thoughts welcome.

I think we should leave the API docs as is as it allows for a ModuleReader to use off-heap memory in the future.

david-beaumont · 2026-01-06T11:23:29Z

src/java.base/share/classes/jdk/internal/jimage/BasicImageReader.java

            }
-
-            ByteBuffer buffer = allocateBuffer(size);
+            ByteBuffer buffer = ByteBuffer.allocate(checkedSize);


I changed this to just "allocate" (not "allocateDirect") after carefully reading up on direct buffers.
It's just possible this might have an observable effect (positive or negative) on either performance or memory pressure, so I'm happy to change it back if people are worried.

It's the only bit of this change that I think might not be a guaranteed no-op.

It avoids allocating direct memory and leaving it to GC reference processing to free, so I think better than just removing the caching.

It should have no impact on usage at runtime, at least not the main stream platforms that are 64-bit and thus memory map the jimage file. On 64-bit, the built-in class loaders and other usages at runtime use slices of the mapping rather rather than allocate.

I think ARM is the only 32-bit port that is still maintained to some degree. 32-bit, and jrtfs when targeting a different run-time image, and the only cases that will use file I/O. I would expect at least some tests for javac -system should exercise this code, and we should have a few jimage tests that run with -Djdk.image.map.all=false to exercise this code (maybe we need more). If you want, you could increase confidence by testing with TEST_OPTS_JAVA_OPTIONS=-Djdk.image.map.all=false. Also since we are early in JDK 27 it means there is time for course correction in case any creepy crawlies come out of the woodwork.

david-beaumont · 2026-01-06T11:27:50Z

src/java.base/share/classes/jdk/internal/jimage/BasicImageReader.java

        }
    }

-    private static ByteBuffer allocateBuffer(long size) {


Since the size check already happened in the caller, this check is never triggered, and if the rounding is removed, the limit() call isn't needed, so this method collapses into a single allocate method call which can be inlined.

RogerRiggs

looks good.

david-beaumont · 2026-01-06T18:19:18Z

/integrate

openjdk · 2026-01-06T18:21:53Z

@david-beaumont
Your change (at version 7d54c8e) is now ready to be sponsored by a Committer.

RogerRiggs · 2026-01-06T19:53:56Z

/sponsor

openjdk · 2026-01-06T19:54:52Z

Going to push as commit 7c979c1.
Since your change was applied there have been 23 commits pushed to the master branch:

53300b4: 8373830: Refactor test/jdk/java/time/test tests to use JUnit over TestNG
f1e0e0c: 8374544: Add SleepyCat diagnostics for all platforms
fbc59ac: 8374555: No need for visible input warning in s.s.u.Password when not reading from System.in
... and 20 more: https://git.openjdk.org/jdk/compare/6ae3e064352a56c5be140fba1ad6d040219432b0...master

Your commit was automatically rebased without conflicts.

openjdk · 2026-01-06T19:55:01Z

@RogerRiggs @david-beaumont Pushed as commit 7c979c1.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

removing ImageBufferCache and unwinding dead code

2b158fe

openjdk bot added the core-libs core-libs-dev@openjdk.org label Jan 5, 2026

openjdk bot added the rfr Pull request is ready for review label Jan 5, 2026

actually remove the class ...

5a37317

david-beaumont commented Jan 5, 2026

View reviewed changes

AlanBateman reviewed Jan 5, 2026

View reviewed changes

david-beaumont added 2 commits January 6, 2026 11:24

copyright update

e3bbc4c

simplifying allocation and argument checks

7d54c8e

david-beaumont commented Jan 6, 2026

View reviewed changes

AlanBateman approved these changes Jan 6, 2026

View reviewed changes

david-beaumont changed the title ~~8374308: ImageBufferCache does not work as intended~~ 8374308: ImageBufferCache has no effect and can be removed Jan 6, 2026

openjdk bot added the ready Pull request is ready to be integrated label Jan 6, 2026

RogerRiggs approved these changes Jan 6, 2026

View reviewed changes

openjdk bot added the sponsor Pull request is ready to be sponsored label Jan 6, 2026

openjdk bot added the integrated Pull request has been integrated label Jan 6, 2026

openjdk bot closed this Jan 6, 2026

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review sponsor Pull request is ready to be sponsored labels Jan 6, 2026

Conversation

david-beaumont commented Jan 5, 2026 • edited by openjdk bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Progress

Issue

Reviewers

Reviewing

Uh oh!

bridgekeeper bot commented Jan 5, 2026

Uh oh!

openjdk bot commented Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openjdk bot commented Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mlbridge bot commented Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Webrevs

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlanBateman Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

RogerRiggs commented Jan 5, 2026

Uh oh!

AlanBateman commented Jan 5, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

RogerRiggs left a comment

Choose a reason for hiding this comment

Uh oh!

david-beaumont commented Jan 6, 2026

Uh oh!

openjdk bot commented Jan 6, 2026

Uh oh!

RogerRiggs commented Jan 6, 2026

Uh oh!

openjdk bot commented Jan 6, 2026

Uh oh!

openjdk bot commented Jan 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

4 participants

david-beaumont commented Jan 5, 2026 •

edited by openjdk bot

Loading

openjdk bot commented Jan 5, 2026 •

edited

Loading

openjdk bot commented Jan 5, 2026 •

edited

Loading

mlbridge bot commented Jan 5, 2026 •

edited

Loading

AlanBateman Jan 5, 2026 •

edited

Loading